Question

我为事件流创建了自定义触发和处理功能。

DataStream<DynamoDBRow> dynamoDBRows =
    sensorEvents
        .keyBy("id")
        .window(GlobalWindows.create())
        .trigger(new MyCustomTrigger())
        .allowedLateness(Time.minutes(1)) # Note
        .process(new MyCustomWindowProcessFunction());

我的触发器基于事件参数。收到事件结束信号后，将 MyCustomWindowProcessFunction（）应用于窗口元素。

@Slf4j
public class MyCustomTrigger extends Trigger<SensorEvent, GlobalWindow> {

  @Override
  public TriggerResult onElement(SensorEvent element, long timestamp, GlobalWindow window, TriggerContext ctx) throws Exception {

    if (element.isEventEnd() == true) {
      return TriggerResult.FIRE_AND_PURGE;
    }

    return TriggerResult.CONTINUE;
  }

  @Override
  public TriggerResult onProcessingTime(long time, GlobalWindow window, TriggerContext ctx) throws Exception {
    return TriggerResult.CONTINUE;
  }

  @Override
  public TriggerResult onEventTime(long time, GlobalWindow window, TriggerContext ctx) throws Exception {
    return TriggerResult.CONTINUE;
  }

  @Override
  public void clear(GlobalWindow window, TriggerContext ctx) throws Exception {}
}

可能很少有传感器数据，即使在触发之后也可能出现。因此，我添加了.allowedLateness(Time.minutes(1))，以确保在处理时不会错过这些事件。

就我而言， allowedLateness 无法正常工作。

浏览文档后，我发现了

如何在GlobalWindow中包含 allowedLateness ？

注意：我还尝试设置环境时间特征

env.setStreamTimeCharacteristic(TimeCharacteristic.IngestionTime);

更新：20-02-2020

目前正在考虑以下方法。（目前无法正常工作）

@Slf4j
public class JourneyTrigger extends Trigger<SensorEvent, GlobalWindow> {

  private final long allowedLatenessMillis;

  public JourneyTrigger(Time allowedLateness) {
    this.allowedLatenessMillis = allowedLateness.toMilliseconds();
  }

  @Override
  public TriggerResult onElement(SensorEvent element, long timestamp, GlobalWindow window, TriggerContext ctx) throws Exception {

    if (element.isEventEnd() == true) {
      log.info("Timer started with allowedLatenessMillis " + allowedLatenessMillis);
      ctx.registerEventTimeTimer(System.currentTimeMillis() + allowedLatenessMillis);
    }

    return TriggerResult.CONTINUE;
  }

  @Override
  public TriggerResult onEventTime(long time, GlobalWindow window, TriggerContext ctx) throws Exception {
    log.info("onEvenTime called at "+System.currentTimeMillis() );
    return TriggerResult.FIRE_AND_PURGE;
  }


  @Override
  public TriggerResult onProcessingTime(long time, GlobalWindow window, TriggerContext ctx) throws Exception {
    return TriggerResult.CONTINUE;
  }

  @Override
  public void clear(GlobalWindow window, TriggerContext ctx) throws Exception {}
}

Answer 1

最后，我可以使用以下自定义触发器满足要求。

import lombok.extern.slf4j.Slf4j;
import org.apache.flink.streaming.api.windowing.time.Time;
import org.apache.flink.streaming.api.windowing.triggers.Trigger;
import org.apache.flink.streaming.api.windowing.triggers.TriggerResult;
import org.apache.flink.streaming.api.windowing.windows.GlobalWindow;

@Slf4j
public class JourneyTrigger extends Trigger<SensorEvent, GlobalWindow> {

  private final long allowedLatenessMillis;

  public JourneyTrigger(Time allowedLateness) {
    this.allowedLatenessMillis = allowedLateness.toMilliseconds();
  }

  @Override
  public TriggerResult onElement(SensorEvent element, long timestamp, GlobalWindow window, TriggerContext ctx) throws Exception {

    if (element.isEventEnd()==true) {
      log.info("Timer started with allowedLatenessMillis " + allowedLatenessMillis);
      ctx.registerProcessingTimeTimer(System.currentTimeMillis() + allowedLatenessMillis);
    }

    return TriggerResult.CONTINUE;
  }

  @Override
  public TriggerResult onProcessingTime(long time, GlobalWindow window, TriggerContext ctx) throws Exception {
    log.info("onProcessingTime called at "+System.currentTimeMillis() );
    return TriggerResult.FIRE_AND_PURGE;
  }

  @Override
  public TriggerResult onEventTime(long time, GlobalWindow window, TriggerContext ctx) throws Exception {
    return TriggerResult.CONTINUE;
  }



  @Override
  public void clear(GlobalWindow window, TriggerContext ctx) throws Exception {}
}

也在Driver.java类中，设置环境时间特征

env.setStreamTimeCharacteristic(TimeCharacteristic.ProcessingTime);

Answer 2

说实话，我看不出在这里使用GlobalWindow的原因。您可以只使用KeyedProcessFunction来达到与Trigger相同的目的，基本上，它将把从事件开始到事件结束的所有元素收集到ListState中然后，当您收到isEventEnd()==true时，只需安排EventTime计时器，该计时器将在一分钟后触发并发出ListState内部收集的结果。

全局窗口自定义触发器上的allowedLateness

2 个答案: